Multiple Comparisons of Return Distributions: A New Look at the Day-of-The-Week Effect
20 Pages Posted: 12 Mar 2002
Date Written: February 28, 2002
Abstract
Day-of-the-week tests for the entire distribution (rather then specific moments) are employed on the Tel-Aviv Stock Exchange index using non-parametric tests. We further add to prior art by examining whether Sunday's return distribution is significantly different from the distribution of any and all other days by employing Multiple Comparison Procedures (MCP). Both techniques, the non-parametric distribution test and the MCP, are not limited to assumptions of distribution normality. The tests were employed on the entire data-set (1988-1998); on three trend-oriented sub-periods (Up-Down-Up), and also conditional on previous day's sign (+/-). We find that Sunday's distribution is closer to Thursdays' (first and last trading days of the week) than to any other day. The MCP reveal significant differences between Sunday's return distribution and Monday-Thursday's for the entire period (1988-1998), and for the sample conditioned on "Down at t-1". The differences are found to a lesser extent for the "Up at t-1" sample, but not at all for the three trend sub-periods. The significant differences were not found for any other trading day.
Keywords: Day of the Week, Multiple Comparison Procedures, Non-Parametric Tests
JEL Classification: C14, C32, G14
Suggested Citation: Suggested Citation
Do you have a job opening that you would like to promote on SSRN?
Recommended Papers
-
Evidence on the Speed of Convergence to Market Efficiency
By Tarun Chordia, Avanidhar Subrahmanyam, ...
-
Equity Returns at the Turn of the Month
By Wei Xu and John J. Mcconnell
-
Testing the Significance of Calendar Effects
By Peter Reinhard Hansen, Asger Lunde, ...
-
The Nepalese Stock Market: Efficiency and Calendar Anomalies
-
Are Calendar Anomalies Still Alive?: Evidence from Istanbul Stock Exchange
By Recep Bildik
-
Do Earnings Explain the January Effect?
By Qingzhong Ma and Hai Lu